hey @Yuri_BC ,
thanks for the connecting impulse! Appreciate that.
I am myself always after keeping discussions organic in forums – and here I am surely the ‘uninitiated’ newbie (which is why I started this as a question 
).
As to the ‘translation’ you are giving, I am not sure, though.
In a way, what I propose is very simple – even if the rationale, I felt, needs some explaining / framing (therefore my long texts).
But: It is not an automated graph as in Obsidian (or InfraNodus, or DiagramGPT,…)– which really is a secondary visualization of what is described on the underlying note / metadata level (– see the argument about secondary visualization vs. primary plane of associative + creative work above; and see @Chr1sG in a way making an argument on the basis of this ‘secondary visualization’ misunderstanding (– if I am allowed to call it a misunderstanding 
–), and my argument for giving the whiteboard and making connections on it some more ‘constitutional’, ‘authorial’ status).
Actually, while being a very visually oriented person when it comes to knowledge work, I like others came to the conclusion that an automated graph view (especially the global ones) doesn´t really bring anything new to the table in most situations.
So, in a sense I am really arguing for the opposite, that is an interactive schematization, visual nevertheless, that can be actively manipulated in the course of knowledge work (I strongly believe in enactivist, constructive knowledge philosophies…). If one can´t interact with it, it is rarely ‘knowledge’… (Here I concur w/ the sentiment voiced by @njyo in the context of one of the graph discussions you referenced: Fibery for knowledge management, what I desire - #3 by njyo )
What I simply expected – and what I want – is a way to register on entity level every instance / act of an entity being placed / created on a whiteboard. This is not what is happening in Obsidian around the graph. And – in difference to the auto-graphs of Obsidian and the like – I am understanding a whiteboard as something that indeed can be constructed and played around with by the users (any idiosyncratic mapping they ‘come up’ with, because of … whatever relevancy they ascribe to the cluster for whatever professional reason).
But really – and this is where a lot of my looong text writing comes from – it is about a kind of 180° swap vis-a-vis things like automated graph representations and generally secondary visualization post factum
– Looking through your links, there is indeed one which really is talking about exactly the same thing. and in a very concise manner; and already in 2021! (– so thank you for making me aware of that! ).
It´s @rothnic who wrote it up in a lucid paragraph:
There are systems/business modeling languages (SysML, BPMN) that implement structured visual modeling concepts that do define explicit and reusable relationships. So, with these systems, you are building the model visually, which is then turned into a database under the hood that can be queried, generate reports, etc. With Fibery and other general-purpose tools you are modeling the relationships and entering the data more directly, then generating visualizations from that database. So it is generally coming at this from the opposite direction.
[ … … just, that I would add: I am also not sure when talking about canvasses, whiteboards etc, we are at all remaining in the conceptual space of ‘graphs’ and ‘trees’ in their stricter logical and technical sense. – This is why I rather use the terms ‘concept map’ and ‘graphical model’, ‘idealized cognitive model’ etc.
This is a discussion much to deep & broad to get into here.… but let me, as a shorthand say / claim, the latter are more attuned / derived from human understanding, they are less strictly ‘logical’ (but organized around – open – domains of interaction and potential / associative linkages, used in complex reasoning tasks rather than in secondary visualization of underlying data structures, they (potentially) accommodate diverse data types, depend on context and conceptual decisions (‘cuts’ / assertions / suppostions),… All in all: what I am more after are associative conceptualizations from the human side feeding into the system, rather than the system giving me derived, secondary visualizations of what is already in there…
like:
… but this would go too far into the field of theory… 
]
Now, if we are talking about Obsidian it is interesting that the largest ‘bang’ it made in the last say 2 years is through the introduction of the Canvas. This is exactly the paradigm shift I am talking about: from automated systems visualization to schematization on a spatial / visual level by the users.
And here we are getting closer to the skin of the game. As the Canvas is a Whiteboard that can be freely and associatively be constructed by the users. It is linked up with / hooked into the ‘underlying’ system of notes and their metadata. Now, the community exactly has a discussion about the question whether and how links drawn on the Canvas – made there always for a reason – should translate back into the core system, especially the backlinks.
In the realm of PKM / Canvas / Notetaking platforms there are different tribes, some believing that all relations from the Canvas should be mirrored in the basic linkage structure; some believing that there should be both ‘visual’ links and harcoded links, somehow…; and some (a minority believes) the links on the Canvas (Whiteboard) basically don´t matter systematically, they are just a visual cue…
You also find this discussion differentiating the orthodoxy and tribes of Scrintal and Heptabase (the other 2 biggies in Canvas-Card-PKMs). Scrintal for hardcoding. Hepta not so much.
So, this discussion really is fully raging in this sector. And having real effects.
I found Fibery, and loved the conceptual avantgarde approach, thus thinking / expecting all that should be possible by choice within it.
Also, I tried to specify some use contexts (maybe almost a dozen up to now) which all have a tally of process logics involved, and thus – with only little elaboration – lead to the ‘real world use cases’ you mention. I just didn´t want to make things longer than they already are, and thought mentioning the scenarios is enough. (Also, I am shying away from really tailoring this to any ‘current product strategy’ … for different reasons: from not being deep in the field adressed (AIUI) – to simply wanting to express my own professional needs, and getting to the ‘insight discovery’ logics involved…)